A Alibaba Cloudcloud computing arm of the Alibaba () group, announced on Wednesday (29) the launch of the Qwen 2.5-Maxnew version of your artificial intelligence model. According to the company, technology surpasses not only DeepSeek-V3, AI model that has shaken the sector in recent days, but also giant solutions such as OpenAi and Meta.
The announcement, made during the Lunar New Year, one of China’s most important holidays, highlights the pressure that DeepSeek’s meteoric growth has been exerting both foreign competitors and Chinese technology companies.
“Qwen 2.5-max widely surpasses the GPT-4O, Deepseek-V3 and Llama-3.1-405B,” said Alibaba Cloud in a statement published in Wechat, referring to the most advanced models of OpenAi and goal.
Deepseek and the AI market dispute
The launch of QWEN 2.5-max happens only 19 days after launching its V3 model and 9 days after the arrival of R1, the latter focused on complex reasoning tasks. Chinese startup innovation surprised the global market, contributing to a strong drop in technology giant actions and raising doubts about billionaire investments in companies such as Microsoft, Google and OpenAi.
The impact of DeepSeek was not restricted to the international scene. Within China itself, the rapid rise of the startup has triggered a race between large companies to update their models.
Tiktok’s controller bytedance was one of the first to react. Just two days after the launch of the DeepSek-R1, the company launched an enhanced version of its AI model, claiming that it surpassed OpenAi O1, in tests that measure the ability of models to understand and perform complex instructions.
Continues after advertising
Alibaba, in turn, had already reduced the prices of its models by up to 97% last year after the launch of DeepSeek-V2, which caused a real price war in China by offering AI services by just 1 yuan (US $ 0.14) for 1 million processed tokens.
(Com Reuters)
